> For the complete documentation index, see [llms.txt](https://docs.hablla.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.hablla.com/hablla-docs-en/automation-flows/all-components/hablla-gpt/autonomous-agent-v3.md).

# Autonomous Agent V3

### Overview

The Autonomous Agent Component V3 (identified internally in the system as `habllaAgentV3`) is an advanced dynamic intelligence and routing node. Its main function is to delegate service to a predefined AI virtual agent, allowing it to conduct fluid conversations with the user autonomously.

What sets this component apart is its architecture of Success-Condition-Based Outputs: the agent operates in interactive loops with the client and the contact only advances in the automation canvas when one of the goals or conditions mapped in the AI robot's tools is fully satisfied, triggering the corresponding routing port.

### Configuration Parameters

The node's parameterization manages robot selection, startup stimulus control, and transactional metadata injection:

* Agent (\* Required Field): Drop-down menu used to select which AI Agent (previously trained and configured on the platform) will take control of the conversation at this stage of the flow.
* Agent skill: Conditional dropdown that defines the specific operational scope or set of specialized tools and instructions (*skills*) that the robot will prioritize during service.
* Enable prompt (Toggle Switch): Controls the conversation startup mode:
  * Unchecked (Disabled): The agent remains passive, waiting and using the last text message sent by the client in the chat to engage in the conversation.
  * Checked (Enabled): Forces the administrator to provide a manual starting directive so the robot can initiate contact proactively.
* Prompt \* (\* Required if key is active): Contextual text box where the greeting instruction or initial operational command is entered. This text functions strictly as the trigger for the agent's first interaction with the client; subsequent messages will follow the natural conversational flow.
* Variables \* (\* Required Field): List-format mapping block (Key and Value) to inject user profile data directly into the AI's context memory (e.g.: `name` and `cellphone`), allowing hyper-personalized interactions.

### Visual Builder

The graphical interface is adaptive and modifies the node's physical morphology in the automation canvas according to the inherited resources:

* Mutable Output Port Mapping: In the flow design editor, this component does not have static generic outputs. As soon as the administrator chooses a *Agent*, the block dynamically renders corresponding side physical ports for each intention or success trigger configured at the AI root (such as outputs for specific department purposes, tickets, or transitions).
* State Flags and Timeout: The node displays small visual control badges at its base for error control (alert sign) and chronological monitoring (clock sign), allowing alternative contingency flows to be connected in case the AI server fails or the client's session expires.

### Special Operators

The logical properties of this component manage conversational session persistence and payload injection:

* Conversational State Retention (Session Standby Loop): Upon entering this node, the traditional linear automation conveyor is paused. The engine creates a persistent session between the AI API and the client's chat. Each new message sent by the user is processed by the language model and responded to autonomously by the robot, without leaving the block, until the transition criterion of one of the output ports is met.
* Context Metadata Interpolation: The repeater fields *Variables* and the box *Prompt* natively support the use of double braces `{{...}}`. This makes it possible to retrieve CRM keys from previous nodes and deliver them formatted for the agent's neural processing.

### Practical Examples

The table below demonstrates traditional scenarios of how to configure the startup and output behavior of Autonomous Agent V3:

| **Selected Agent** | **Key "Enable prompt"** | **Initial Prompt Configuration**                                                                             | **Available Outputs in the Node**                                                                               | **Practical Use Case**                                                                                                                                                                                                                               |
| ------------------ | ----------------------- | ------------------------------------------------------------------------------------------------------------ | --------------------------------------------------------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| AI Reception       | Disabled                | *None (Starts with user input)*                                                                              | <p><code>support</code></p><p><br></p><p><code>finance</code></p><p><br></p><p><code>broker</code></p>          | The customer sends a vague message in the DM. The AI Agent takes over, talks freely with them and, as soon as it identifies that the customer's desire is to buy a property, dispatches the contact through the physical output `broker`.            |
| Sales Agent        | Enabled                 | "Hello `{{name}}`! I noticed you abandoned the cart for the plan. Would you like to clear up any questions?" | <p><code>purchased</code></p><p><br></p><p><code>declined</code></p><p><br></p><p><code>human</code></p>        | The component is used in an active time-based flow. The AI actively starts the conversation with the prompt phrase. If it convinces the customer, it exits through `purchased`; if the customer asks to speak to a person, it exits through `human`. |
| General Screening  | Disabled                | *None*                                                                                                       | <p><code>send\_images</code></p><p><br></p><p><code>scheduled</code></p><p><br></p><p><code>fallback</code></p> | The bot talks with the client to collect documents for a registration. As soon as the user sends the requested files, the AI validates receipt and routes the contact through the success output `send_images` to continue through the flow.         |


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:

```
GET https://docs.hablla.com/hablla-docs-en/automation-flows/all-components/hablla-gpt/autonomous-agent-v3.md?ask=<question>&goal=<endgoal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.